Since 1996 when Highly Pathogenic Avian Influenza type H5N1 first emerged in southern China, numerous studies sought risk factors and produced risk maps based on environmental and anthropogenic predictors. However little attention has been paid to the link between the level of intensification of poultry production and the risk of outbreak. This study revised H5N1 risk mapping in Central and Western Thailand during the second wave of the 2004 epidemic. Production structure was quantified using a disaggregation methodology based on the number of poultry per holding. Population densities of extensively- and intensively-raised ducks and chickens were derived both at the sub-district and at the village levels. LandSat images were used to derive another previously neglected potential predictor of HPAI H5N1 risk: the proportion of water in the landscape resulting from floods. We used Monte Carlo simulation of Boosted Regression Trees models of predictor variables to characterize the risk of HPAI H5N1. Maps of mean risk and uncertainty were derived both at the sub-district and the village levels. The overall accuracy of Boosted Regression Trees models was comparable to that of logistic regression approaches. The proportion of area flooded made the highest contribution to predicting the risk of outbreak, followed by the densities of intensively-raised ducks, extensively-raised ducks and human population. Our results showed that as little as 15% of flooded land in villages is sufficient to reach the maximum level of risk associated with this variable. The spatial pattern of predicted risk is similar to previous work: areas at risk are mainly located along the flood plain of the Chao Phraya river and to the south-east of Bangkok. Using high-resolution village-level poultry census data, rather than sub-district data, the spatial accuracy of predictions was enhanced to highlight local variations in risk. Such maps provide useful information to guide intervention.

Global concerns over the emergence of zoonotic pandemics emphasize the need for high-resolution population distribution mapping and spatial modelling. Ongoing efforts to model disease risk in China have been hindered by a lack of available species level distribution maps for poultry. The goal of this study was to develop 1 km resolution population density models for China’s chickens, ducks, and geese. We used an information theoretic approach to predict poultry densities based on statistical relationships between poultry census data and high-resolution agro-ecological predictor variables. Model predictions were validated by comparing goodness of fit measures (root mean square error and correlation coefficient) for observed and predicted values for ¼ of the sample data which was not used for model training. Final output included mean and coefficient of variation maps for each species. We tested the quality of models produced using three predictor datasets and 4 regional stratification methods. For predictor variables, a combination of traditional predictors for livestock mapping and land use predictors produced the best goodness of fit scores. Comparison of regional stratifications indicated that for chickens and ducks, a stratification based on livestock production systems produced the best results; for geese, an agro-ecological stratification produced best results. However, for all species, each method of regional stratification produced significantly better goodness of fit scores than the global model. Here we provide descriptive methods, analytical comparisons, and model output for China’s first high resolution, species level poultry distribution maps. Output will be made available to the scientific and public community for use in a wide range of applications from epidemiological studies to livestock policy and management initiatives.

Background

Ecological niche modeling integrates known sites of occurrence of species or phenomena with data on environmental variation across landscapes to infer environmental spaces potentially inhabited (i.e., the ecological niche) to generate predictive maps of potential distributions in geographic space. Key inputs to this process include raster data layers characterizing spatial variation in environmental parameters, such as vegetation indices from remotely sensed satellite imagery. The extent to which ecological niche models reflect real-world distributions depends on a number of factors, but an obvious concern is the quality and content of the environmental data layers.

Methods

We assessed ecological niche model predictions of H5N1 avian flu presence quantitatively within and among four geographic regions, based on models incorporating two means of summarizing three vegetation indices derived from the MODIS satellite. We evaluated our models for predictive ability using partial ROC analysis and GLM ANOVA to compare performance among indices and regions.

Results

We found correlations between vegetation indices to be high, such that they contain information that overlaps broadly. Neither the type of vegetation index used nor method of summary affected model performance significantly. However, the degree to which model predictions had to be transferred (i.e., projected onto landscapes and conditions not represented on the landscape of training) impacted predictive strength greatly (within-region model predictions far out-performed models projected among regions).

Conclusion

Our results provide the first quantitative tests of most appropriate uses of different remotely sensed data sets in ecological niche modeling applications. While our testing did not result in a decisive "best" index product or means of summarizing indices, it emphasizes the need for careful evaluation of products used in modeling (e.g. matching temporal dimensions and spatial resolution) for optimum performance, instead of simple reliance on large numbers of data layers.

Highly pathogenic avian influenza (HPAI) H5N1 was first encountered in 1996 in Guangdong province (China) and started spreading throughout Asia and the western Palearctic in 2004–2006. Compared to several other countries where the HPAI H5N1 distribution has been studied in some detail, little is known about the environmental correlates of the HPAI H5N1 distribution in China. HPAI H5N1 clinical disease outbreaks, and HPAI virus (HPAIV) H5N1 isolated from active risk-based surveillance sampling of domestic poultry (referred to as HPAIV H5N1 surveillance positives in this manuscript) were modeled separately using seven risk variables: chicken, domestic waterfowl population density, proportion of land covered by rice or surface water, cropping intensity, elevation, and human population density. We used bootstrapped logistic regression and boosted regression trees (BRT) with cross-validation to identify the weight of each variable, to assess the predictive power of the models, and to map the distribution of HPAI H5N1 risk. HPAI H5N1 clinical disease outbreak occurrence in domestic poultry was mainly associated with chicken density, human population density, and elevation. In contrast, HPAIV H5N1 infection identified by risk-based surveillance was associated with domestic waterfowl density, human population density, and the proportion of land covered by surface water. Both models had a high explanatory power (mean AUC ranging from 0.864 to 0.967). The map of HPAIV H5N1 risk distribution based on active surveillance data emphasized areas south of the Yangtze River, while the distribution of reported outbreak risk extended further North, where the density of poultry and humans is higher. We quantified the statistical association between HPAI H5N1 outbreak, HPAIV distribution and post-vaccination levels of seropositivity (percentage of effective post-vaccination seroconversion in vaccinated birds) and found that provinces with either outbreaks or HPAIV H5N1 surveillance positives in 2007–2009 appeared to have had lower antibody response to vaccination. The distribution of HPAI H5N1 risk in China appears more limited geographically than previously assessed, offering prospects for better targeted surveillance and control interventions.

Author Summary

The geographical distribution of highly pathogenic avian influenza (HPAI) H5N1 and agro-ecological risk factors have been studied in a number of countries in Southeast Asia. However, little is know of its distribution in China where HPAI H5N1 first emerged in 1996, evolved, and spread throughout Asia and the western Palearctic in 2004–2006. This study analyzes separately the distribution, in domestic poultry, of HPAI virus (HPAIV) H5N1 isolated from active risk-based surveillance sampling and HPAI H5N1 clinical disease outbreaks. These data are analyzed in relation to the distribution of chicken and domestic waterfowl population density, proportion of land covered by rice or surface water, cropping intensity, elevation, and human population density. HPAI H5N1 viruses identified by risk-based surveillance are found to be associated with domestic waterfowl density, human population density, and the proportion of land covered by surface water. In contrast, HPAI H5N1 clinical disease outbreak occurrences were mainly associated with chicken density, human population density, and low elevation. These results show that the distribution of HPAI H5N1 risk in China appears more limited geographically than previously assessed, offering prospects for better targeted surveillance and control interventions.

Highly pathogenic avian influenza (HPAI) H5N1 virus persists in Asia, posing a threat to poultry, wild birds, and humans. Previous work in Southeast Asia demonstrated that HPAI H5N1 risk is related to domestic ducks and people. Other studies discussed the role of migratory birds in the long distance spread of HPAI H5N1. However, the interplay between local persistence and long-distance dispersal has never been studied. We expand previous geospatial risk analysis to include South and Southeast Asia, and integrate the analysis with migration data of satellite-tracked wild waterfowl along the Central Asia flyway. We find that the population of domestic duck is the main factor delineating areas at risk of HPAI H5N1 spread in domestic poultry in South Asia, and that other risk factors, such as human population and chicken density, are associated with HPAI H5N1 risk within those areas. We also find that satellite tracked birds (Ruddy Shelduck and two Bar-headed Geese) reveal a direct spatio-temporal link between the HPAI H5N1 hot-spots identified in India and Bangladesh through our risk model, and the wild bird outbreaks in May–June–July 2009 in China (Qinghai Lake), Mongolia, and Russia. This suggests that the continental-scale dynamics of HPAI H5N1 are structured as a number of persistence areas delineated by domestic ducks, connected by rare transmission through migratory waterfowl.

Electronic supplementary material

The online version of this article (doi:10.1007/s10393-010-0672-8) contains supplementary material, which is available to authorized users.

Beginning in 2003, highly pathogenic avian influenza (HPAI) H5N1 virus spread across Southeast Asia, causing unprecedented epidemics. Thailand was massively infected in 2004 and 2005 and continues today to experience sporadic outbreaks. While research findings suggest that the spread of HPAI H5N1 is influenced primarily by trade patterns, identifying the anthropogenic risk factors involved remains a challenge. In this study, we investigated which anthropogenic factors played a role in the risk of HPAI in Thailand using outbreak data from the “second wave” of the epidemic (3 July 2004 to 5 May 2005) in the country. We first performed a spatial analysis of the relative risk of HPAI H5N1 at the subdistrict level based on a hierarchical Bayesian model. We observed a strong spatial heterogeneity of the relative risk. We then tested a set of potential risk factors in a multivariable linear model. The results confirmed the role of free-grazing ducks and rice-cropping intensity but showed a weak association with fighting cock density. The results also revealed a set of anthropogenic factors significantly linked with the risk of HPAI. High risk was associated strongly with densely populated areas, short distances to a highway junction, and short distances to large cities. These findings highlight a new explanatory pattern for the risk of HPAI and indicate that, in addition to agro-environmental factors, anthropogenic factors play an important role in the spread of H5N1. To limit the spread of future outbreaks, efforts to control the movement of poultry products must be sustained.

Outbreaks of highly pathogenic avian influenza (HPAI) H5N1 subtype have occurred in many countries across Asia, Europe, and Africa since 2003. Better understanding of the ecology and risk factors of HPAI is critical for surveillance, risk assessment, and public health policy. We introduce satellite remote sensing as one important tool, and highlight the potential of using satellite images to monitor dynamics of climate and landscapes that are related to wild bird migration and agriculture in the context of avian influenza transmission.

Highly pathogenic avian influenza (HPAI) caused by H5N1 viruses has become a global scale problem which first emerged in southern China and from there spread to other countries in Southeast and East Asia, where it was first confirmed in end 2003. In previous work, geospatial analyses demonstrated that free grazing ducks played critical role in the epidemiology of the disease in Thailand in the winter 2004/2005, both in terms of HPAI emergence and spread. This study explored the geographic association between free grazing duck census counts and current statistics on the spatial distribution of rice crops in Thailand, in particular the crop calendar of rice production. The analysis was carried out using both district level rice statistics and rice distribution data predicted with the aid of remote sensing, using a rice-detection algorithm. The results indicated a strong association between the number of free grazing ducks and the number of months during which second-crop rice harvest takes place, as well as with the rice crop intensity as predicted by remote sensing. These results confirmed that free grazing duck husbandry was strongly driven by agricultural land use and rice crop intensity, and that this later variable can be readily predicted using remote sensing. Analysis of rice cropping patterns may provide an indication of the location of populations of free grazing ducks in other countries with similar mixed duck and rice production systems and less detailed duck census data. Apart from free ranging ducks and rice cropping, the role of hydrology and seasonality of wetlands and water bodies in the HPAI risk analysis is also discussed in relation to the presumed dry season aggregation of wild waterfowl and aquatic poultry offering much scope for virus transmission.

Anatids may have spread the virus along their autumn migration routes.

During the second half of 2005, highly pathogenic avian influenza (HPAI) H5N1 virus spread rapidly from central Asia to eastern Europe. The relative roles of wild migratory birds and the poultry trade are still unclear, given that little is yet known about the range of virus hosts, precise movements of migratory birds, or routes of illegal poultry trade. We document and discuss the spread of the HPAI H5N1 virus in relation to species-specific flyways of Anatidae species (ducks, geese, and swans) and climate. We conclude that the spread of HPAI H5N1 virus from Russia and Kazakhstan to the Black Sea basin is consistent in space and time with the hypothesis that birds in the Anatidae family have seeded the virus along their autumn migration routes.